Anecdotal uncertainty of pain — LessWrong